search

jnimmo / hass-intesishome

Experimental fork of the IntesisHome integration for Home Assistant
19 stars 11 forks source link

Can't regain connection after relatively long time without HA restart. #3

Open Jimmypl007 opened 2 years ago

Jimmypl007 commented 2 years ago

I mentioned this error in my last comment on the other issue I raised (although I'm not sure if that last comment has been seen) I had hoped the update would have fixed it or perhaps the tweak you instructed to reduce connection times, but nope..

This is possibly directly related to my other issue of the constant connection losses, but after a stretch of days '10 or more' of connection dropouts and reconnects it's like it just randomly gives up (a restart seems to restart the time/days it takes before it stops trying anymore)

this screenshot shows the graph of connections and then at the end they give up. while it might not necessarily mean it takes a larger stretch of days this hasn't happened unless I haven't restarted for a good week.

also ( i assume ) due to the dropout and failures to connect, the log is spammed with errors seen here and here now I'm not sure if the runner.py thing is related but it's the only thing that appears to error around as much as the intesis thing.

jnimmo commented 2 years ago

If it's giving up after a number of days, there is likely an unhandled exception killing the loop, which ie something we'll need to look into.

Connection reset by peer however implies Intesis may be doing some kind of rate limiting (restricting how long a given IP address can be connected to the port for). Recent versions are sending keepalives so I'm not sure what else can do for that particular issue.

May have to consider adding an alternative connection option which just uses HTTP polling every 60 seconds or something rather than trying to maintain a persistent connection.

Jimmypl007 commented 2 years ago

Hmm, I get a general idea but a fair bit goes right over my head... As usual, if you need any log information or tests done, let me know.

Jimmypl007 commented 1 year ago

Thought I will add I have noticed this problem occurs after smaller periods of time as well. not just multiple days. the last one lasted 23 hours. Although I understand this doesn't change the likelihood of "an unhandled exception"